熊猫-如何将不同的功能应用于df.series而该功能取决于另一个系列？

Question

Simply speaking, I want to do this:

df.s1.map["a":df.s2+1, "b":df.s2-1]

However,it won't work because df.s2 is not a constant.

How can I map s1 on df.s2+1 while s1='a' ,and map s1 on df.s2-1 while s1='b' ?

Answer 1

You can use apply using lambda with axis = 1 ie

if you have a dataframe

df = pd.DataFrame({'s1':['a','b','a','a','b'],'s2':[1,3,5,2,4]})
df['s3'] = df.apply(lambda x : x['s2']+1 if x['s1'] == 'a'  else x['s2']-1 ,axis=1 )

Or take d['s2'] common from the dictionary and then add it at the end after mapping.

df['s3'] = df.s1.map({'a':1,'b':-1})+df['s2']

Output:

s1  s2  s3
0  a  1   2 
1  b  3   2 
2  a  5   6 
3  a  2   3 
4  b  4   3 
In [364]:

Output :